Perfusion of His-Tagged Eukaryotic Myocilin Increases Outflow Resistance in Human Anterior Segments in the Presence of Aqueous Humor

Abstract

purpose. A previous study by the authors has shown that recombinant myocilin purified from a prokaryotic expression system increases outflow resistance in cultured human anterior segments. The present study was performed to determine whether full-length myocilin purified from a human trabecular meshwork cell expression system alters outflow resistance after infusion into human anterior segments.

methods. A feline immunodeficiency virus vector encoding both full-length myocilin (amino acids 1-503 fused to C-terminal V5 and six-histidine epitopes) and puromycin resistance was used to transduce a transformed trabecular meshwork cell line (TM5). Stably expressing cells were selected with puromycin. Recombinant myocilin was purified from the media using nickel ion affinity chromatography. Control purifications were performed on media from parental TM5 cells. Anterior segments of human eyes were placed in organ culture and perfused with either Dulbecco’s modified Eagle’s medium (DMEM) or DMEM supplemented with 50% porcine aqueous humor. One eye received an anterior chamber exchange with recombinant myocilin (2 μg/mL), whereas the fellow eye received an equal volume of control. Immunohistochemistry was performed with anti-myocilin and anti-V5 antibodies. Native polyacrylamide gel electrophoresis was used to analyze myocilin complex formation in porcine aqueous humor.

results. Recombinant myocilin in porcine aqueous humor increased outflow resistance in cultured human anterior segments (91% ± 68% [mean ± SD] versus 18% ± 31% in fellow control eye; n = 9, P = 0.004). Maximum outflow resistance was obtained 5 to 17 hours after infusion and remained above baseline for >3 days. Recombinant myocilin also increased outflow resistance in eyes incubated in DMEM, but only if myocilin was preincubated with porcine aqueous humor (78% ± 77% when preincubated in DMEM containing porcine aqueous humor versus 13% ± 15% when preincubated with DMEM alone, n = 6, P = 0.03). Recombinant myocilin appears to form a complex in porcine aqueous humor with a heat-labile protein(s). Immunohistochemistry revealed the presence of myocilin in the juxtacanalicular region of the trabecular meshwork.

conclusions. Myocilin purified from human trabecular meshwork cells increased outflow resistance in cultured human anterior segments, but only after incubation with porcine aqueous humor. Recombinant myocilin appears to form a complex in porcine aqueous humor that enables it to bind specifically within the trabecular meshwork.

Myocilin is a glycoprotein found in aqueous humor and expressed in many ocular tissues including the trabecular meshwork. ^{1 2 3 4 5 6 7 8 9 10 11} First identified as a protein that was induced in trabecular meshwork cells after steroid treatment, ^{12 13} myocilin has more recently been identified as a glaucoma-associated protein after mutations of the gene were found to be associated with autosomal dominant forms of juvenile open-angle glaucoma. ^{14 15} Over 60 mutations have been identified in the myocilin gene, and these mutations are associated with approximately 8% of juvenile open-angle glaucoma and 4% of primary open-angle glaucoma cases. ^{15 16 17 18}  

The function of myocilin is unknown, although evidence supports both an intracellular and extracellular role. Intracellularly, myocilin binds to photoreceptors in the retina ¹ and to myosin regulatory light chain, a component of the myosin motor protein. ¹⁹ Myocilin has also been found to associate with mitochondria. ²⁰ More evidence supports an extracellular role for myocilin. Myocilin contains an N-terminal signal peptide sequence characteristic of secreted proteins. Studies have shown that myocilin can bind to extracellular molecules such as fibronectin, optimedin, and several forms of collagen. ^{19 21 22 23} It can also bind to itself through interactions within an N-terminal leucine zipper. ²⁴ In the C-terminal region, myocilin contains homology to a family of glycosylated extracellular matrix proteins called olfactomedins. Most glaucoma-associated mutations are found in the olfactomedin homology domain. 

Although myocilin mutations have been associated with juvenile open-angle glaucoma and primary open-angle glaucoma, evidence exists that abnormally high levels of normal myocilin may be involved in obstructing the aqueous outflow pathway. In primary open-angle glaucoma, increased levels of myocilin have been observed in the trabecular meshwork of some eyes. ²⁵ Experimentally, intraocular pressure and myocilin levels increase in a time-dependent manner after corticosteroid treatment in human perfusion organ culture, mimicking corticosteroid induced ocular hypertension. ²⁶ Furthermore, we have shown that the infusion of full-length recombinant myocilin partially purified from a prokaryotic expression system increased outflow resistance in perfusion organ culture. ²⁷ The infusion of only the C-terminal portion of myocilin (containing the olfactomedin homology region) did not change outflow resistance, ²⁸ suggesting that myocilin may have to be full-length to alter the aqueous outflow pathway. 

We describe a eukaryotic expression system that produces full-length myocilin. Purified recombinant eukaryotically expressed myocilin was infused into cultured human anterior segments that were cultured with or without porcine aqueous humor supplementation. Changes in outflow resistance were evaluated. 

Methods

Myocilin Construct and Stable Cell-Line Generation

Myocilin cDNA ²⁷ containing base pairs 30 to 1551 (GenBank accession no., U85257; http://www.ncbi.nlm.nih.gov/Genbank; provided in the public domain by the National Center for Biotechnology Information, Bethesda, MD) was amplified by PCR and inserted into expression plasmid pcDNA 4 V5/His (Invitrogen, Carlsbad, CA) generating pcDNA4.myoc. A CMV-promoter/human myocilin cassette was amplified by PCR from plasmid pcDNA4.myoc and inserted into pGINWF ^{29 30} between XhoI and EcoRI, generating the plasmid pTINWF2. This cassette contains a CMV-promoter 5′ of the full-length myocilin cDNA (codes for amino acids 1-504). Because selection of pTINWF2 (internal ribosomal entry site [IRES]-mediated expression of neomycin-phosphotransferase) in transduced TM5 cells (transformed trabecular meshwork cells; gift from Abe Clark, Alcon Inc., Fort Worth, TX) was inefficient, the IRES-neomycin phosphotransferase resistance cassette was replaced with the IRES-puromycin resistance cassette from plasmid pEF/IRES.puro ³¹ by inserting EcoRI-IRES.puro-Mfe1 into pTINWF2 to replace EcoRI-IRES.neo-EcoRI. 

A second myocilin expression vector was made by amplifying myocilin cDNA from base pairs 30 to 1548 (no termination codon) and inserting it into pcDNA 4 V5/His, thus generating expression plasmid pcDNA4.myocV5his. This expression vector contains the full coding region of myocilin (no termination codon) fused to the V5/histidine-tag (pcDNA4.myocV5his). The 3′ two thirds of this vector was PCR amplified to introduce a 3′ EcoRI site and used to replace Bsu361-myoc-EcoRI in pTINWF2, yielding plasmid pTV5hisIPWF. 

In summary, pTV5hisIPWF is a feline immunodeficiency virus (FIV) transfer vector plasmid ³² with internal CMV promoter-mediated expression of full-length human myocilin fused with V5/histidine epitopes and cap-independent, IRES-mediated coexpression of puromycin resistance for selection of transduced cells. The cloned myocilin cDNA including 3′ V5 and histidine tags in pTV5hisIPWF was sequenced to confirm absence of any inadvertent mutations. 

To generate FIV vectors, the following plasmids were transfected into 293T cells: TV5hisIPWF (described earlier), pFP93 (packaging plasmid), and pMD-G (expresses vesicular stomatitis virus glycoprotein G; VSV-G). ^{30 33} Medium was replaced 12 to 16 hours later, and vector supernates were collected 48 hours thereafter. Media were filtered (0.45 μm) and concentrated by ultracentrifugation, as previously described. ³⁴ Transducing units (TDU/mL) were determined by incubating Crandell feline kidney (CrFK) cells ³⁵ with increasing dilutions of FIV-myocilin vector in the presence of puromycin (2 μg/mL) and counting puromycin-resistant colonies. 

For production of stable TM5 cells expressing myocilin, 1 × 10⁸ TDU of FIV vector containing full-length myocilin tagged with V5/histidine epitopes was added to approximately 300,000 TM5 cells. After a 6-hour incubation at 37°C, the vector was removed, and fresh Dulbecco’s modified Eagle’s medium (DMEM; Cellgro, Herndon, VA) containing 10% fetal bovine serum (DMEM-FBS) was added to the cells. Forty-eight hours after transduction, medium was removed and fresh DMEM-FBS with puromycin (2 μg/mL) was added. TM5 cells expressing myocilin were selected in puromycin for 1 month. Stable cells (TM5-myocilin) were combined and maintained as a polyclonal cell population in DMEM-FBS. 

Purification of Myocilin

Conditioned medium (45 mL) from TM5-myocilin cells was mixed with 4.5 mL of 10× preparation buffer (0.5 M NaH₂PO₄ [pH 8.0], 1.5 M NaCl, 0.1 M imidazole). Ni-NTA resin (4 mL; Qiagen, Valencia, CA) was added and the sample shaken at 4°C for 2 hours. After incubation, Ni-NTA resin was pelleted by centrifugation at 2500g for 5 minutes, resuspended in 40 mL of wash buffer (50 mM NaH ₂PO₄ [pH 8.0], 300 mM NaCl, 10 mM imidazole), and incubated with shaking at 4°C for 10 minutes. Resin was repelleted at 2500g for 5 minutes, and the supernatant was removed. The Ni-NTA resin was washed five times. Two additional washes were performed with 50 mM NaH₂ PO₄ (pH 8.0), 300 mM NaCl, and 50 mM imidazole. For elution of myocilin from the nickel, the resin was resuspended in 4 mL of 50 mM NaH₂ PO₄ (pH 8.0), 300 mM NaCl, 250 mM imidazole and shaken at 4°C for 10 minutes. Nickel resin was centrifuged at 2500g for 5 minutes, and the supernatant was isolated. The elution step was repeated. The two elutions were combined and dialyzed overnight (minimum of 16 hours) against buffer containing 50 mM NaH₂ PO₄ (pH 8.0), 300 mM NaCl, and decreasing amounts of imidazole (100 mM→50 mM→0 mM). Four different purification batches were used in this study. 

A separate, identical purification was performed on 3-day conditioned medium from parental TM5 cells. The medium was manipulated in the same way and at the same time as that described for myocilin purification. This purification was performed to control for potential effects on outflow resistance of low levels of nonspecific, copurifying proteins. 

Analysis of Purified Myocilin

Samples of the recombinant myocilin purification process (30 μL each) were placed in Laemmli buffer, boiled, and separated on 4% to 15% SDS-PAGE gradient gels (Bio-Rad, Hercules, CA). Gels were either stained blue (GelCode Blue; Pierce, Rockford, IL), or the proteins were transferred to polyvinylidene diflouride membrane (Millipore, Billerica, MA) in 1× transfer buffer (50 mM Tris, 384 mM glycine, 0.01% SDS, and 20% methanol). 

For Western blot analysis, membranes were blocked in 20 mM Tris (pH 7.5), 150 mM NaCl, 0.05% Tween, and 2% evaporated milk, and probed with either a rabbit polyclonal myocilin antibody or a mouse monoclonal anti-V5 antibody (Invitrogen) and followed with either a secondary horseradish-peroxidase–linked anti-rabbit or anti-mouse Ig antibody (GE Healthcare, Piscataway, NJ). Antibody-antigen complexes were detected using ECL Western blot signal detection reagent (GE Healthcare). 

Porcine Aqueous Humor

Aqueous humor was collected from pigs within 1 hour of death at Applied Pork Products, Inc. (Austin, MN). A 27-gauge needle attached to a tuberculin syringe was inserted into the anterior chamber, and the aqueous was slowly withdrawn. Aspiration was stopped before the anterior chamber collapsed to minimize contamination of primary aqueous humor. Aqueous humor was immediately placed on ice until all samples were collected and then centrifuged twice at 4000g to remove pigment and debris material, aliquoted, and stored at −70°C. 

Human Organ Cultures

Normal human eyes were obtained within 12 hours of death from the Minnesota Lions Eye Bank and placed in anterior segment perfusion culture. Eighteen pairs of eyes from 11 male and 7 female donors with an average age of 73 ± 9 years (mean ± SD; range: 57–91 years) were used in three separate studies. No eyes had glaucoma or uveitis or were from patients on topical eye medications. The culture technique was similar to that described previously. ^{36 37} Eyes were bisected at the equator, and the iris, lens, and vitreous were removed. The anterior segment was clamped in a modified Petri dish and the eye perfused with either DMEM (Cellgro) with added antibiotics (penicillin: 10,000 units; streptomycin: 10 mg; amphotericin B: 25 mg; and gentamicin: 1.7 mg in 100 mL medium; standard culture medium referred to as DMEM in text) or DMEM with antibiotics supplemented with porcine aqueous humor (50% final concentration; referred to as DMEM-AH in the text). Porcine aqueous humor was used in place of human aqueous humor, because it is more readily available in the amounts necessary to perform the experiments. We have previously reported that the addition of porcine aqueous humor to the anterior segment of the human eye has improved trabecular cell viability and maintained molecular characteristics that are more similar to the trabecular meshwork of fresh human eyes (Fautsch MP, et al. IOVS 2003;44:ARVO E-Abstract 3165). Anterior segments were perfused at the normal human flow rate (2.5 μL/min). The anterior segments were cultured at 37°C in a 5% CO₂ atmosphere. Intraocular pressures were continuously monitored with a pressure transducer connected to the second access cannula built into the dish and recorded with an automated computerized system. 

Study 1: Anterior Segments Perfused with DMEM-AH.

Twelve pairs of anterior segments were perfused with DMEM-AH. Recombinant myocilin (2 μg) was mixed with 500 μL of porcine aqueous humor, diluted with DMEM to 1 mL, and incubated at 4°C for 2 hours (n = 9). Purified control (equal volume to recombinant myocilin) was also mixed with 500 μL of porcine aqueous humor, diluted with DMEM to 1 mL, and incubated at 4°C for 2 hours (n = 9). Incubations at 4°C were performed to reduce any chance of myocilin degradation. Three additional pairs of anterior segments were tested, using a 37°C incubation for 2 hours, to prove that potential aggregate formation that may occur in the cold was not responsible for changes in outflow resistance. 

After incubation, one anterior segment received recombinant myocilin, and the fellow anterior segment received purified control by anterior chamber exchange. The anterior chamber exchanges were performed with a gravity-driven, constant-pressure method over a 5-minute period. Pressure data from either eye were not used until the end of the first hour after the anterior chamber exchange. 

Study 2: Anterior Segments Perfused with DMEM.

To determine whether aqueous was necessary for the effect of myocilin on outflow resistance, three pairs of anterior segments were perfused with DMEM (did not contain porcine aqueous humor). Both anterior segments were infused with recombinant myocilin after a 2-hour incubation at 4°C in the presence or absence of porcine aqueous humor. One eye received a 1-mL infusion of porcine aqueous humor (500 μL) containing recombinant myocilin (2 μg; diluted to 500 μL with DMEM). The fellow eye received an infusion of recombinant myocilin (2 μg) diluted to 1 mL with DMEM (contained no porcine aqueous humor). Four days after the first infusion, a second infusion was performed, but in crossover fashion, using identical samples. The eye that received recombinant myocilin in porcine aqueous humor in the first infusion received recombinant myocilin in DMEM (no porcine aqueous humor) in the second infusion. The fellow eye that received recombinant myocilin in DMEM in the first infusion received recombinant myocilin in porcine aqueous humor in the second infusion. Anterior exchanges were performed as described in study 1. 

Study 3: Anterior Segments Perfused with DMEM and Human Albumin.

This was done for two reasons: (1) to determine whether the albumin in aqueous humor interacts with myocilin to cause a change in outflow resistance, and (2) to determine whether the change in resistance associated with myocilin is specific for myocilin, or could be caused by any similar-sized protein (the albumin by itself in the control eye). Three pairs of anterior segments were perfused with DMEM containing 25 μg/mL human albumin (Sigma-Aldrich, St. Louis, MO). This dose was chosen to represent the albumin concentration in normal human aqueous (12–50 μg/μL ^{38 39} ; albumin is 50% of total protein concentration). Recombinant myocilin (2 μg) or purified control proteins were incubated for 2 hours at 4°C in DMEM containing 25 μg/mL human albumin. One eye received a 1-mL infusion of recombinant myocilin, DMEM, and human albumin, and the fellow eye received a 1-mL infusion of purified control proteins, DMEM, and human albumin. Anterior exchanges were performed as described in study 1. 

Preservation of Tissue for Immunohistochemistry

All eyes in this study were analyzed by light microscopy to assess trabecular meshwork appearance. Two tissue samples (180° apart) were dissected from each eye and fixed in 4% paraformaldehyde (in phosphate buffer; pH 8.0), dehydrated in a graded series of ethanol (75%, 85%, 95%, and 100%) and embedded in paraffin. Sections 5-μm-thick were mounted on glass slides (Superfrost/Plus; Fisher, Pittsburgh, PA) and baked at 60°C for 2 hours. 

For immunohistochemistry, tissue sections were deparaffinized in xylene and rehydrated in a graded series of ethanol (100%, 95%, 80%, and 70%). Sections were incubated at 95°C in 1 mM EDTA (pH 8.0), as previously described. ²⁷ Tissue sections were blocked in phosphate-buffered saline containing 3% bovine serum albumin and 0.1% Triton X-100 and probed with either rabbit anti-myocilin antibody ²⁷ or mouse monoclonal anti-V5 antibody (Invitrogen). FITC-conjugated anti-rabbit Ig or Alexa Fluor 594 goat anti-mouse Ig was used as the secondary antibody (Molecular Probes, Eugene, OR). Tissue was also stained with 4′,6′-diamino-2-phenylindole (DAPI) to highlight the nuclei (0.25 μg/mL). 

Analysis of Myocilin Incubated in Porcine Aqueous Humor

To identify complex formation, recombinant myocilin (final concentration 2 μg/mL) was incubated at 4°C for 2 hours in DMEM alone, DMEM-AH, or heat-denatured DMEM-AH (porcine aqueous humor and DMEM heated at 80°C for 10 minutes before recombinant myocilin was added). After incubation, 15 μL of each sample was separated on either a native 4% to 15% gradient PAGE gel (nondenatured, nonreduced) or a 4% to 15% gradient SDS-PAGE gel (denatured, reduced). Proteins were transferred to a polyvinylidene diflouride membrane (Millipore) in 1× transfer buffer and probed with mouse monoclonal anti-V5 antibody (Invitrogen). Horseradish-peroxidase–linked anti-mouse Ig (GE Healthcare) was used as secondary antibody. Antibody–antigen complexes were detected using ECL Western blot signal detection reagent (GE Healthcare). 

Coimmunoprecipitation of Porcine Albumin and Recombinant Myocilin

Recombinant myocilin (2 μg) or purified control proteins (previously described) were diluted to 500 μL with DMEM and mixed with 500 μL porcine aqueous humor in duplicate. One set of reactions (recombinant myocilin and purified control) was incubated at 4°C (2 hours) and the second set of reactions was incubated at 37°C (2 hours). After incubation, 250 μL of each reaction was mixed with 500 μL PBS in triplicate. One set of reactions received no antibody, one set received 10 μL anti-myocilin antibody, and the third set received 10 μL anti-porcine albumin antibody (Bethyl Laboratories, Montgomery, TX). Samples were incubated overnight at 4°C. After incubation, 75 μL of a protein A-Sepharose/PBS slurry (1:1) was added to each sample and incubated at 4°C for 75 minutes. Samples were centrifuged at 13,000 rpm for 1 minute, the supernatant discarded, and the Sepharose pellet was resuspended in 900 μL phosphate-buffered saline. Samples were mixed for 2 minutes and centrifuged again. This process was repeated two additional times. After completion of washing, Sepharose pellets were resuspended in Laemmli sample buffer, boiled, and separated on a 4% to 15% gradient gel. Gel was transferred to PVDF membrane and probed with anti-porcine albumin or anti-myocilin antibody. 

Statistics

Outflow facility (C) was calculated as F/P where F is the flow rate (2.5 μL/min) and P is the pressure (mm Hg). Resistance was calculated as 1/C. Change in resistance was calculated according to the formula (R _experimental/R _baseline − 1) × 100. Outflow resistances for each eye of the experimental group or the control group were combined and the mean calculated. Comparisons of recombinant myocilin-infused eye to control group were analyzed with either a signed-rank or a paired t-test. 

Results

Purification of Eukaryotically Expressed Recombinant Myocilin

Initial attempts to express and purify recombinant myocilin from baculovirus (high-five cells) and HEK293 cells resulted in a cleaved myocilin product. Although some full-length myocilin was present, most of the myocilin had been cleaved at amino acid 214 (based on protein sequence analysis), generating a C-terminal fragment of approximately 32 kDa (Fig. 1A) . 

In contrast, trabecular cells in monolayer culture are known to express myocilin as a full-length protein with minimal cleavage. ^{26 40} We hypothesized that trabecular cells possessed a mechanism that enabled expression and secretion of full-length myocilin. To test this, we made a stable cell line by transducing a transformed trabecular cell line (TM5, discussed in the Methods section) with full-length myocilin fused to C-terminal V5 and six histidine epitopes. Western blot analysis verified expression of full-length myocilin and also showed that myocilin was secreted as a doublet, similar to in vivo myocilin found in human trabecular meshwork and aqueous humor (Fig. 1B) . In contrast to the HEK293 expression system, only a small proportion of myocilin was cleaved from the TM5-myocilin stable cell line. A stained gel showed that the purification procedure was highly specific, showing only the myocilin doublet (Fig. 1C) . Protein sequence analysis verified that the purified protein was myocilin. 

Analysis of Outflow Resistance after Infusion of Recombinant Myocilin

To test the hypothesis that excess wild-type myocilin purified from a trabecular cell expression system could affect outflow resistance in the human eye, nine pairs of human anterior segments were cultured in DMEM-AH and infused with 2 μg of recombinant myocilin or an equal volume of purified control (purification of conditioned medium from parental TM5 cells) that had been preincubated at 4°C. A 2-μg dose of recombinant myocilin was used because this dose is comparable to that used in our previous study in which recombinant myocilin partially purified from a prokaryotic expression system increased outflow resistance by 94% in human anterior segments. ²⁷  

The baseline outflow facilities were similar between anterior segments receiving recombinant myocilin and their fellow control eyes (0.14 ± 0.04 vs. 0.15 ± 0.06; mean ± SD). Infusion with recombinant myocilin resulted in a steady increase in IOP that reached maximum levels within 5 to 17 hours (Fig. 2A) . IOP remained above baseline for 72 to 96 hours. In all nine eyes infused with recombinant myocilin, outflow resistance was increased above control (Fig. 2B) . At 9 hours, anterior segments infused with myocilin had a mean increase in outflow resistance of 91% ± 68%, whereas the fellow control eye increased 18% ± 31% (mean ± SD; P = 0.004, experimental versus control). Significant increases were also seen at 12 and 24 hours after infusion with recombinant myocilin. 

To verify that the change in outflow resistance seen in anterior segments infused with recombinant myocilin was not due to nonphysiologic aggregates that formed during incubation at 4°C, we performed incubations at 37°C. Analysis of eyes infused with recombinant myocilin or purified control proteins that were preincubated with DMEM-AH at 37°C showed results similar to those obtained at 4°C. At 9, 12, and 24 hours after infusion, eyes that received recombinant myocilin all increased outflow resistance above control (9 hours after myocilin infusion: 67% ± 43%; control: 34% ± 37%; n = 3; due to the small sample size, significance was not calculated). 

Location of Recombinant Myocilin in the Trabecular Meshwork

To determine the location of recombinant myocilin in the trabecular meshwork after infusion, immunohistochemistry was performed. At maximum pressure, recombinant myocilin was found mainly in the juxtacanalicular region, just below and extending the length of Schlemm’s canal (Fig. 3A) . Recombinant myocilin was also seen in the uveal and corneoscleral regions of the meshwork in lesser amounts. A myocilin-specific antibody labeled both the exogenous and the endogenous myocilin (Fig. 3B) . DAPI staining showed numerous trabecular cell nuclei, a good indicator of successful cultures (Fig. 3C) . A composite image of the trabecular meshwork shows the localization of recombinant with endogenous myocilin in trabecular cells (Fig. 3D) . Immunohistochemistry showed the specificity of the antibodies in the fellow control eye and verified the health of the cells, including numerous intact nuclei (Figs. 3E 3F 3G 3H) . 

If the quantity and location of recombinant myocilin is important for increasing outflow resistance in anterior segment culture, changes in the amount and location of recombinant myocilin should be seen in the trabecular meshwork as outflow resistance returns to baseline. Immunohistochemistry performed on eyes with an initial increase in resistance after myocilin addition followed by a gradual decrease to near baseline levels at 90 hours showed markedly less recombinant myocilin in the trabecular meshwork (Fig. 4) . In comparison to its presence in eyes analyzed at maximum pressure (Fig. 3) , only a small amount of recombinant myocilin was present in the juxtacanalicular region. This suggests that the change in outflow resistance correlates with the quantity and location of recombinant myocilin. 

Porcine Aqueous versus DMEM

To determine the role of porcine aqueous humor in the myocilin-induced increase in outflow resistance, a crossover experiment was performed. Three pairs of eyes were placed in perfusion organ culture with DMEM (did not contain porcine aqueous humor). In one eye, recombinant myocilin (2 μg/mL) was added after incubation in DMEM-AH for 2 hours at 4°C. In the fellow eye, recombinant myocilin (2 μg/mL) was added after incubation in DMEM (no porcine aqueous humor supplement) for 2 hours at 4°C. Eyes were maintained in culture for an additional 4 to 5 days. At this time, the crossover was performed, as the initial eye receiving recombinant myocilin incubated in DMEM-AH was switched to recombinant myocilin in DMEM (no porcine aqueous humor). The fellow eye that initially received recombinant myocilin incubated in DMEM now received recombinant myocilin incubated in DMEM-AH. A representative pressure graph is shown in Figure 5A . 

Eyes infused with recombinant myocilin incubated in aqueous had a significantly greater increase in resistance than the control eyes receiving myocilin in DMEM. When the crossover infusions were performed, recombinant myocilin preincubated in DMEM-AH also increased resistance in the former control eye, which had not responded to myocilin preincubated in DMEM in each of the three pairs of eyes. The overall outflow resistance change at 9 hours in all eyes infused with myocilin preincubated in DMEM-AH was 78% ± 77%, compared with 13% ± 15% for the eyes infused with myocilin preincubated with DMEM (P = 0.03; n = 6; Fig. 5B ). Similar trends were seen at 12 and 24 hours after infusion, suggesting that recombinant myocilin might be forming a complex with a constituent of porcine aqueous humor which then obstructed outflow. 

Analysis of Recombinant Myocilin in Porcine Aqueous Humor

To determine whether myocilin associates with a protein(s) in porcine aqueous humor, we analyzed recombinant myocilin, with and without porcine aqueous humor, by native polyacrylamide gel electrophoresis. In the presence of porcine aqueous humor, recombinant myocilin formed complexes that migrated differently from those formed in DMEM (Fig. 6 , compare lanes 1 and 2). Furthermore, heat-denaturation of porcine aqueous humor before incubation with recombinant myocilin disrupted this complex (Fig. 6 , compare lane 2 and 3), suggesting that the complex formation of myocilin in porcine aqueous humor is dependent on a heat-labile protein(s). 

Role of Albumin in Myocilin-Induced Increase in Outflow Resistance

Myocilin and albumin have both been implicated in the obstruction of flow through microporous filters. ^{6 41} To test whether albumin is the factor associated with myocilin that causes an increase in outflow resistance, we infused one eye with recombinant myocilin that had been incubated with DMEM containing human albumin (25 μg/mL) while DMEM containing human albumin alone was delivered to the fellow eye (n = 3). Two of the three eyes that received recombinant myocilin did not show any change in outflow resistance. A third pair showed an increase in outflow resistance, but most of the increase occurred within the first hour after infusion, which was a different pattern than seen for recombinant myocilin incubated in porcine aqueous humor. This suggests that albumin combined with recombinant myocilin is not sufficient to alter outflow resistance in a consistent manner. Furthermore, incubation of recombinant myocilin with porcine aqueous humor followed by immunoprecipitation with porcine albumin or myocilin-specific antibodies, did not show interaction of these two proteins at either 4°C or 37°C (Fig. 7) . This further suggests that albumin is probably not the binding partner necessary for myocilin-induced outflow increase in resistance. 

Discussion

Increased levels of myocilin have been associated with corticosteroid induced ocular hypertension and primary open-angle glaucoma. ^{12 13 25 26} Our study supports the idea that excess myocilin can increase outflow resistance in the human eye. Infusion of recombinant myocilin increased outflow resistance significantly more than the control, and the change in pressure remained above baseline levels for 3 to 4 days. The presence of myocilin in the juxtacanalicular region at maximum pressure supports the importance of this region in outflow resistance. Furthermore, recombinant myocilin appears to exert its effect on the aqueous outflow pathway only after it complexes with a protein(s) in porcine aqueous humor. 

Our finding that purified myocilin produced from trabecular cells can increase outflow resistance in human anterior segments is consistent with our previous study in which we used recombinant myocilin produced from a prokaryotic expression system. ²⁷ Both studies showed that infusion of excess myocilin increases outflow resistance. Although the result was similar, differences between the two studies are evident. In the present study, myocilin was expressed, maintained, and purified from conditioned medium as a full-length protein, showing a doublet near 58 kDa (the C-terminal tag adds 28 amino acids). This is similar to in vivo myocilin that is N-glycosylated and is a doublet with molecular mass between 53 and 57 kDa. In the previous study, prokaryotically expressed myocilin was purified from whole-cell lysate as a full-length protein, but was not glycosylated, because bacteria lack the organelles essential for carbohydrate addition to proteins. Although eukaryotically expressed myocilin did not change outflow resistance in human anterior segments incubated in DMEM, prokaryotically expressed myocilin increased outflow resistance under similar conditions. Differences in the glycosylation patterns or in the processing of these proteins between the expression systems may alter how these proteins interact with themselves or other interacting proteins, which may result in the functional difference in outflow resistance between the recombinant myocilin produced from eukaryotic and prokaryotic expression systems. 

Another difference is that myocilin produced from eukaryotic cells required porcine aqueous humor to induce changes in outflow resistance. Aqueous humor is the normal nutrient source for the anterior tissues of the eye. Its inclusion in the medium of anterior segments in culture improves trabecular cell viability and also maintains molecular characteristics that more closely represent the fresh, noncultured eye (Fautsch MP, et al. IOVS 2003;44:ARVO E-Abstract 3165). Components of aqueous have been implicated in the regulation of the aqueous outflow pathway by studies showing that aqueous humor can bind and reduce flow through polycarbonate filters. ^{6 28} The ability of recombinant myocilin to form a complex in vitro that alters the aqueous outflow pathway further suggests the importance of aqueous humor and its components in normal fluid flow through the trabecular meshwork. The present study shows that albumin, the predominant protein found in aqueous humor, is probably not myocilin’s primary binding partner for altering outflow resistance. Further studies are warranted to identify the protein(s) in aqueous that interacts with myocilin. 

The recombinant myocilin used in our studies was fused to a C-terminal tag containing a six-histidine epitope for purification by nickel ion affinity chromatography and an epitope for the V5 monoclonal antibody. Although the addition of a C-terminal tag to myocilin is a potential limitation, current purification methods have not been described for native myocilin. Many recombinant proteins containing epitope tags have maintained their function when compared with native protein. However, because the function of myocilin has not been determined, the effects of the C-terminal tag on myocilin function and binding of associated proteins is unknown. 

Other studies have not associated myocilin with changes in outflow resistance. ^{28 42} Infusion of a myocilin fragment containing the olfactomedin homology domain (amino acids 215-503 fused to a histidine epitope; purified from HEK293 cells) did not increase outflow resistance in human anterior segments. ^{28 42} This fragment lacked the N-terminal region that contains the leucine zipper, a region known to be important for myocilin dimerization. ²⁴ Second, the infusion of the C-terminal region of myocilin was performed in eyes incubated in DMEM. In our study, only full-length myocilin incubated in porcine aqueous humor caused an increase in outflow resistance. 

Transgenic mice expressing myocilin from a lens-specific βB1-crystallin promoter show increased levels of myocilin in aqueous humor but no change in outflow resistance in young adult mice. ⁴² Myocilin levels in the aqueous humor of these mice was calculated at 0.2 μg/mL, which is 10 times lower than our infusion of recombinant myocilin (2 μg/mL). The concentration of myocilin in aqueous humor has not been reported and it is unknown whether physiologic levels of myocilin are as high as those we infused into human anterior segments. Effects of perfused myocilin are lasting, causing intraocular pressure elevation for at least 3 days. If an increase in secretion occurred in vivo, it could accumulate in the trabecular meshwork over an extended period. Nevertheless, the ability of excess myocilin to increase outflow resistance supports the theory that an increase in myocilin may interfere with the normal hydrodynamics within the trabecular meshwork, resulting in changes in the aqueous outflow pathway. 

The species and age of the eyes were also different in our infusion studies and the transgenic mouse study. In our study, we used anterior segments from human donors averaging 73 ± 9 years (mean ± SD, n = 18). The use of eyes from the elderly is important because age-related changes, particularly due to loss of trabecular cells and elasticity of the trabecular beams, may effect how the aqueous outflow pathway functions. Excess levels of myocilin may only contribute to the elevated outflow resistance in eyes in which age-related changes are already present. In comparison, the myocilin transgenic mice were analyzed at 12 weeks of life, which is the equivalent to a young adult human. Further investigation, particularly as the mice age, will be useful for testing the theory that increased levels of myocilin alter the aqueous outflow pathway. 

 Supported in part by National Eye Institute Grant EY07065; Research to Prevent Blindness, Inc.; the philanthropy of The Bonner Glaucoma Research Fund; and the Mayo Foundation.

 Submitted for publication March 15, 2005; revised August 8 and September 20, 2005; accepted November 17, 2005.

 Disclosure: M. P. Fautsch, None; C. K. Bahler, None; A. M. Vrabel, None; K. G. Howell, None; N. Loewen, None; W. L. Teo, None; E. M. Poeschla, None; D.H. Johnson, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Purification of recombinant eukaryotic myocilin. (A) Western blot of 3-day conditioned medium collected from stable HEK293 cells expressing myocilin and probed with anti-V5 antibody. (B) Western blot analysis of 3-day conditioned medium collected from stable TM5-myocilin cells. Blots were probed with anti-V5 or myocilin-specific antibodies. Samples represent various stages of the myocilin purification process. Lane 1: 3-day conditioned medium isolated from TM5-myocilin cells. Lane 2: 250 mM imidazole elution of purified myocilin. Lane 3: purified myocilin after dialysis. Lane 4: nickel-purified 3-day conditioned medium from parental TM5 cells after dialysis. Myocilin is secreted as a doublet, although in Lane 1 the migration of the two proteins is compressed due to the large quantity of albumin in fetal bovine serum. (C) Blue-stained duplicate of the gel in (B). Purified myocilin from TM5-myocilin expressing cells remains mostly full length. Note the large amount of 32-kDa fragment of myocilin from a HEK293 stable cell line (A) in comparison to TM5-myocilin-expressing cells (B).

Figure 2.

 

Purified recombinant myocilin increases outflow resistance. (A) Pressure graph indicating an increase in IOP 9 hours after addition of 2 μg/mL of purified myocilin in DMEM-AH (arrow). The fellow eye received an equal volume of control purification in DMEM-AH. (B) Outflow resistance in eye pairs in which the anterior segment was incubated in DMEM-AH and infused with DMEM-AH containing purified recombinant myocilin (2 μg/mL) or fellow anterior segment incubated with DMEM-AH and infused with DMEM-AH containing the control. The graph in (A) is represented in (B) by eye set 6. *Signed-rank test significance for myocilin versus the control (P = 0.004).

Figure 3.

 

Immunohistochemistry of purified myocilin at maximum outflow resistance. Paraffin-embedded sections of anterior segments cultured in DMEM-AH 17 hours after infusion of purified myocilin preincubated in DMEM-AH. (A–D) Infused with purified myocilin. Arrows: location of recombinant eukaryotic myocilin. (E–H) Infused with control. (A, E) Probed with anti-V5 antibody. (B, F) Probed with anti-myocilin antibody. (C, G) DAPI stained. (D) Combination of (A–C). (H) Combination of (E–G).

Figure 4.

 

Immunohistochemistry of purified myocilin 90 hours after infusion. Paraffin-embedded sections of anterior segments cultured in DMEM-AH 90 hours after infusion of purified recombinant myocilin preincubated in DMEM-AH. Pressure remained above baseline for 60 hours after myocilin infusion and gradually lowered to near baseline levels at 90 hours. (A–D) Infused with purified myocilin. Arrow indicates location of residual levels of recombinant myocilin. (E–H) Infused with control. (A, E) Probed with anti-V5 antibody. (B, F) Probed with anti-myocilin antibody. (C, G) DAPI stained. (D) Combination of (A–C). (H) Combination of (E–G).

Figure 5.

 

Purified recombinant myocilin increases outflow resistance after incubation in DMEM-AH but not in DMEM alone. (A) Representative pressure graph of one pair of eyes cultured in DMEM indicating an increase in IOP 9 hours after infusion of purified myocilin (2 μg/mL) that was preincubated with DMEM-AH. The fellow eye received purified myocilin (2 μg/mL) that was preincubated in DMEM alone. A crossover experiment was performed at 108 hours. (B) Three sets (1, 2, and 3) of eyes in which two experimental conditions were tested: (a) One eye received purified myocilin (2 μg/mL) incubated in DMEM-AH, and the fellow eye received purified myocilin (2 μg/mL) incubated in DMEM alone; (b) same conditions as for the pair of eyes in (a) except the opposite eyes were infused 4 to 5 days after the initial infusion. *Signed-rank test significance for myocilin in DMEM-AH versus myocilin in DMEM (P = 0.03). Graph in (A) is represented in (B) by eye set 3a and 3b.

Figure 6.

 

Myocilin complexes in aqueous humor. Western blot of native PAGE gel or SDS-PAGE gel probed with anti-V5 antibody. Myocilin forms a complex that migrates differently on native-PAGE gels after incubation for 2 hours at 4°C in DMEM-AH (lane 2) when compared with purified myocilin incubated with DMEM alone (lane 1) or heat-denatured (HD) DMEM-AH (lane 3). Lanes 5 to 8: equal loads of recombinant myocilin from the same reaction that was used on the native gel. The anti-V5 antibody recognizes a nonspecific 55- to 60-kDa protein under native conditions in porcine aqueous humor. This band is not recognized in heat-denatured porcine aqueous humor or on SDS-PAGE.

Figure 7.

 

Myocilin and porcine albumin do not interact. Western blot probed with (A) porcine albumin or (B) V5 after immunoprecipitation with no antibody (control), myocilin antibody (lane 1), or porcine albumin antibody (lane 2). Recombinant myocilin or purified control proteins were incubated in DMEM-AH for 2 hours at either 4°C or 37°C before immunoprecipitation. No albumin signal in the myocilin antibody immunoprecipitation reactions (lane 2 of the recombinant myocilin/DMEM-AH reaction) suggests that myocilin and albumin are not protein-binding partners. The presence of myocilin is confirmed in this lane in (B).